EUVIMEDEUVIMED entdecken
Crossref

Systematic Construction of Continuous-Time Neural Networks for Linear Dynamical Systems

Chinmay Datar, Adwait Datar, Felix Dietrich, Wil Schilders

SIAM Journal on Scientific Computing · 31.08.2025 · 47 (4) · C820-C845

DOI: 10.1137/24m1648946 ↗

Kurz gefasst

Fachartikel in der Zeitschrift SIAM Journal on Scientific Computing, online erschienen am 10.07.2025. Verfasst von Chinmay Datar und 3 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 78 Literaturangaben. Gefördert von: Institute for Advanced Study at the Technical University of Munich, Deutsche Forschungsgemeinschaft, Nederlandse Organisatie voor Wetenschappelijk Onderzoek.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Discovering a suitable neural network architecture for modeling complex dynamical systems poses a formidable challenge, often involving extensive trial and error and navigation through a high-dimensional hyperparameter space. In this paper, we discuss a systematic approach to constructing neural architectures for modeling a subclass of dynamical systems, namely, linear time-invariant (LTI) systems. We use a variant of continuous-time neural networks in which the output of each neuron evolves continuously as a solution of a first-order or second-order ordinary differential equation. Instead of deriving the network architecture and parameters from data, we propose a gradient-free algorithm to compute sparse architecture and network parameters directly from the given LTI system, leveraging its properties. We bring forth a novel neural architecture paradigm featuring horizontal hidden layers and provide insights into why employing conventional neural architectures with vertical hidden layers may not be favorable. We also provide an upper bound on the numerical errors of our neural networks. Finally, we demonstrate the high accuracy of our constructed networks on three numerical examples. Reproducibility of computational results. This paper has been awarded the “SIAM Reproducibility Badge: Code and data available” as a recognition that the authors have followed reproducibility principles valued by SISC and the scientific computing community. Code and data that allow readers to reproduce the results in this paper are available at https://gitlab.com/chinmay_datar/dynamic-neural-networks/-/tree/main/dynamic_neural_networks/numerical_examples and in the supplementary materials ( dynamic-neural-networks-main.zip [1.94MB]), ( M164894_Supplement.pdf [2.05MB]). [Formula: see text]

Förderung

  • Institute for Advanced Study at the Technical University of Munich
  • Deutsche Forschungsgemeinschaft
  • Nederlandse Organisatie voor Wetenschappelijk Onderzoek

Zitieren

APA

Datar, C., Datar, A., Dietrich, F., & Schilders, W. (2025). Systematic Construction of Continuous-Time Neural Networks for Linear Dynamical Systems. SIAM Journal on Scientific Computing, 47(4), C820-C845. https://doi.org/10.1137/24m1648946

Vancouver

Datar C, Datar A, Dietrich F, Schilders W. Systematic Construction of Continuous-Time Neural Networks for Linear Dynamical Systems. SIAM Journal on Scientific Computing. 2025;47(4):C820-C845. doi:10.1137/24m1648946

BibTeX
@article{Datar2025,
  author = {Datar, Chinmay and Datar, Adwait and Dietrich, Felix and Schilders, Wil},
  title = {Systematic Construction of Continuous-Time Neural Networks for Linear Dynamical Systems},
  journal = {SIAM Journal on Scientific Computing},
  year = {2025},
  volume = {47},
  number = {4},
  pages = {C820--C845},
  doi = {10.1137/24m1648946}
}
Kostenlos starten →